1. Technical Field
The present invention relates to an image reading apparatus such as a scanner that reads images of original documents.
2. Related Art
Typical image reading apparatuses separate a plurality of original documents mounted on an original platen (placement section) one by one, transport the original documents, and read images of the original documents (see, for example, JP-A-2012-41118).
Such image reading apparatuses include a multi-sheet feed detection sensor (multi-sheet feed detection section) for detecting whether a plurality of original documents are transported in an overlapped state (hereinafter, also referred to as a multi-sheet feed). Such image reading apparatuses have a separation mode for separating the original documents on the original platen and transporting the original documents, and a non-separation mode for manually feeding the original documents. While the non-separation mode is selected, the multi-sheet feed state detected by the multi-sheet feed detection sensor is disabled.
In other words, in the image reading apparatuses, original documents such as envelopes and center-folded original documents, which are transported one by one but may be erroneously detected as a multi-sheet feed, are manually fed one by one in the non-separation mode.
In some cases, however, image reading apparatuses have to read a plurality of original documents including original documents such as envelopes that tend to be erroneously detected as a multi-sheet feed, even if a single envelope is fed, and original documents that are rarely detected as an error such as paper. In such a case, it is laborious to feed the original documents one by one to read the original documents in the non-separation mode. On the other hand, if the original documents are read in the separation mode, the multi-sheet feed detection sensor may erroneously detect an error, and the reading operation may be stopped. In such a case, the erroneously detected original documents are to be read in the non-separation mode, and this operation causes the read data to be split into a plurality of files corresponding to before and after the detected error.